Methotrexate (MTX) is a disease-modifying antirheumatic drug (DMARD) used for the treatment of rheumatoid arthritis (RA). Genetic polymorphisms in AMP deaminase (AMPD1), 5-aminoimidazole-4-carboxamide ribonucleotide formyltransferase (ATIC), inosine triphosphate pyrophosphorylase (ITPA) and methylenetetrahydrofolate dehydrogenase (MTHFD1) were shown to be associated with disease activity, MTX treatment response and MTX-induced toxicity in patients with RA.1,–,6 The aim of our study was to test these associations in Slovenian patients with RA treated with MTX.

Our study included 211 patients with RA who were previously characterised.7 Among them, 98 patients were receiving MTX monotherapy, 57 co-treated with one or more DMARD and 56 discontinued MTX before enrolment owing to MTX inefficacy and/or toxicity. The 28-joint count Disease Activity Score (DAS28) of two consecutive visits was assessed. DAS28 ≤3.2 was defined as low and DAS28 ≫3.2 as moderate or high disease activity.8 Adverse drug reactions (ADRs) were recorded from patients' files to evaluate MTX toxicity. All the patients were Central European Caucasian, and gave their written informed consent before enrolment.

TaqMan SNP genotyping assays (C_33603912_10, C_16218146_10 and C_1376137_10; ABI, Foster City, California, USA) were used to determine AMPD1 C34T (rs17602729), ATIC C347G (rs2372536) and MTHFD1 G1958A (rs17850560) polymorphisms, respectively. A custom TaqMan single nucleotide polymorphism (SNP) genotyping assay with the following primers and probes, F: 5′-GGAGGTCGTTCAGATTCTAGGAGAT-3′, R: 5′-TCCGGCAGGTCAATTTTCTGT-3′ and VIC-AAGTGCATGTAAACTT, FAM-TGCATGGAAACTT was used to determine ITPA A94C (rs1127354) polymorphism. Binary logistic regression with the addition of gender, disease and MTX treatment duration, MTX dose and the presence of rheumatoid factor and anti-citrullinated peptide antibodies was built using SPSS 14.0.1 software (SPSS, Chicago, Illinois, USA).

In the study group, 34 patients (16.1%) had low disease activity, 177 (83.9%) had moderate/high disease activity and 146 (69.2%) had MTX-induced ADRs. The observed frequencies of investigated polymorphisms are presented in table 1. Carriers of AMPD1 34T allele had 3.8-fold higher probability for low disease activity compared with non-carriers (p=0.012, odds ratio (OR)=3.786, 95% CI 1.347 to 10.642). No patients with ITPA 94A allele and low disease activity were found. ATIC C347G and MTHFD1 G1958A polymorphisms were not significantly associated with disease activity; however, carriers of ATIC 347G allele had 2.5-fold higher risk for MTX-induced toxicity than non-carriers (p=0.024, OR=2.497, 95% CI 1.129 to 5.525). After exclusion of patients co-treated with other DMARDs, we observed even stronger association of the AMPD1 34T allele with disease activity (p=0.006, OR=6.729, 95% CI 1.741 to 26.007) and found significant association of the MTHFD1 1958GG genotype with low disease activity (p=0.021, OR=4.674, 95% CI 1.266 to 17.262).

Recently, a clinical pharmacogenetic model of MTX efficacy identified AMPD1 C34T, ATIC C347G, ITPA A94C and MTHFD1 G1958A polymorphisms as predictors of good clinical response at 6 months of treatment.6 Although we could not validate this model owing to the retrospective nature of our data, we confirmed the association of AMPD1 34T and MTHFD1 1958G alleles with low disease activity and of ATIC 347G allele with MTX toxicity, similar to previous reports.2 4,–,6 The data on association of ATIC C347G polymorphism and MTX treatment response are, however, inconsistent, as both ATIC 347CC and GG genotypes were associated with a better clinical response.1 3 5

In conclusion, our results support the view that pharmacogenetic data may help to predict the efficacy and safety of MTX treatment of RA in clinical practice.